摘要:
Techniques are provided for managing caches in a system with multiple caches that may contain different copies of the same data item. Specifically, techniques are provided for coordinating the write-to-disk operations performed on such data items to ensure that older versions of the data item are not written over newer versions, and to reduce the amount of processing required to recover after a failure. Various approaches are provided in which a master is used to coordinate with the multiple caches to cause a data item to be written to persistent storage. Techniques are also provided for transferring data items and locks associated with the data items from one node to another.
摘要:
Described herein are techniques for generating a global checkpoint system change number and computing a snapshot query using the global checkpoint system change number without a need to acquire global locks. In many cases, the need to acquire global locks is eliminated, thereby saving the overhead attendant to processing global locks.
摘要:
Techniques are provided for managing caches in a system with multiple caches that may contain different copies of the same data item. Specifically, techniques are provided for coordinating the write-to-disk operations performed on such data items to ensure that older versions of the data item are not written over newer versions, and to reduce the amount of processing required to recover after a failure. Various approaches are provided in which a master is used to coordinate with the multiple caches to cause a data item to be written to persistent storage. Techniques are also provided for managing checkpoints associated with the caches, where the checkpoints are used to determine the position at which to begin processing recovery logs in the event of a failure.
摘要:
Methods and apparatuses for providing access to data in a cluster database are described. Embodiments of the invention include identifying situations in which many nodes require relatively equal access to a data block, and taking appropriate measures to improve the performance of the system in these situations. For example, a node may automatically downgrade its lock from an X lock to an S lock as soon as it has finished a modification of the data. S locks are sent to other nodes in the system as soon as a modification is complete. Modifications to the data are also sent to other nodes in the system as soon as a modification is complete. Depending on the size of the data modification, the entire data block may be sent, or only the modifications to the data may be sent.
摘要:
A method for preventing multiple pings. An embodiment of the invention detects requests of data blocks entailing pings likely to cause additional pings. The servicing of requests involving a pings likely to cause additional pings is deferred until a service enabling conditions occurs. Another embodiment of the invention detects situations where by further updating a data block before pinging the data block use of resources on the remote node requesting the data block are reduced. The servicing of the request for the data block is deferred until a service enabling conditions occurs.
摘要:
Described herein are techniques for generating a global checkpoint system change number and computing a snapshot query using the global checkpoint system change number without a need to acquire global locks. In many cases, the need to acquire global locks is eliminated, thereby saving the overhead attendant to processing global locks.
摘要:
Techniques are provided for managing caches in a system with multiple caches that may contain different copies of the same data item. Specifically, techniques are provided for coordinating the write-to-disk operations performed on such data items to ensure that older versions of the data item are not written over newer versions, and to reduce the amount of processing required to recover after a failure. Various approaches are provided in which a master is used to coordinate with the multiple caches to cause a data item to be written to persistent storage. Techniques are also provided for managing checkpoints associated with the caches, where the checkpoints are used to determine the position at which to begin processing recovery logs in the event of a failure.
摘要:
Techniques are provided for performing changes to a resource governed by a locking mechanism. An entity (such as a server instance in a database system cluster) requests permission to modify the resource. In response to the request, the entity receives a first lock on the resource, which grants permission to perform the change to the resource without making the change permanent. After receiving the first lock, the entity performs the change to a copy of the resource that resides in shared memory without making another copy of the resource. After performing the change and until receiving permission to make the change permanent, the entity prevents the change to the resource from becoming permanent. After performing the change, the entity receives a second lock on the resource, which grants the entity permission to make the change permanent. After receiving the second lock, the entity ceases to prevent the change to the resource from becoming permanent.
摘要:
A lock manager system is provided that includes multiple distinct mechanisms for assigning resources to masters. Specifically, the lock manager system includes (1) one or more assignment mechanisms that do not take into account the affinity between nodes and objects (“affinity-independent assignment mechanisms”), and (3) one or more assignment mechanisms that do take into account the affinity between nodes and objects (“affinity-based assignment mechanisms”). At any given time, the assignment of any given resource to a master is governed by only one of the multiple assignment mechanisms employed by the system. Various techniques are described for determining which assignment mechanism to use with which resources.
摘要:
Techniques are provided for managing caches in a system with multiple caches that may contain different copies of the same data item. Specifically, techniques are provided for coordinating the write-to-disk operations performed on such data items to ensure that older versions of the data item are not written over newer versions, and to reduce the amount of processing required to recover after a failure. Various approaches are provided in which a master is used to coordinate with the multiple caches to cause a data item to be written to persistent storage. Techniques are also provided for transferring data items and locks associated with the data items from one node to another.